AbstentionBench - answer unknown - KUQ/Future unknowns - Recall: leaderboard

Metric: Abstention Recall (%). Source: github.com. 20 models tracked.

Top models

#ModelScore
1Llama-3.1-Tulu-3-70B-DPO81.6
2Llama-3.1-Tulu-3-70B80.06
3Gemini 1.5 Pro75.77
4Llama-3.1-Tulu-3-8B-DPO73.62
5Llama-3.1-Tulu-3-8B73.31
6GPT-4o71.17
7Llama 3.1 8B Instruct69.94
8Llama 3.3 70B Instruct69.94
9Mistral 7B Instruct (v0.3)68.4
10Qwen 2.5 32B Instruct66.56
11O166.56
12Llama 3.1 405B Instruct65.34
13Llama-3.1-Tulu-3-8B-SFT62.58
14Llama 3.1 70B Instruct62.27
15DeepSeek R1 Distill Llama 70B54.29

Interactive version: theaggregate.ai/benchmark?slug=abstentionbench-answer-unknown-kuq-future-unknowns-recall · How It Works · Data refreshed daily, snapshot 2026-09-19.